Guo Weixing

I have the same problem and I would like to be able to use normalized predicted values and normalized target values when calculating losses. But the normal output is unnormalized...

> So can you check dataloader .dataset.data['reals'], this should be normalized. The input seems to be normalized, but the normalized y (target vector) is not found in the data. ```...

> What is the dimension of your sale_cols?  These are some of the time-sliding statistics generated based on the 'sale' column.

> I think the reals should already contain the target as well, since you include it as unknown_reals.  In line 476 of timeseries.py, self._preprocess_data(data) returns a dataframe in which,...